Google sues SerpAPI, signaling war on the AI data supply chain

Google has officially declared war on the data brokers powering the next generation of AI search, filing a major lawsuit against SerpAPI for what it calls "unlawful scraping" and the brazen circumvention of security measures.

The suit, announced by Google General Counsel Halimah DeLaine Prado, alleges that SerpAPI uses "shady back doors"—including cloaking, massive bot networks, and constantly changing identities—to steal copyrighted and licensed content directly from Google Search results. This content, which includes images from Knowledge Panels and real-time data from Search features, is then allegedly resold by SerpAPI for a fee, willfully disregarding the rights of the original content owners.

This is not a standard cease-and-desist. Google is asking a court to issue an injunction to stop SerpAPI’s operations entirely, framing the action as a necessary defense of the web ecosystem and the content providers who rely on Google’s protocols.

For years, the web scraping industry has operated in a legal gray zone, often protected by precedents that affirm the right to collect publicly accessible data. But Google’s move, coupled with recent events, suggests that the gray zone is rapidly shrinking, replaced by a hard, expensive line drawn by deep-pocketed incumbents.

The most immediate and chilling precedent is the abrupt shutdown of Proxycurl in July 2025, which StartupHub.ai reported exclusively. Proxycurl, a successful data enrichment API focused on LinkedIn profiles, closed shop after LinkedIn and Microsoft filed a sweeping lawsuit. While Proxycurl insisted it only scraped public data, the lawsuit alleged the company relied on "hundreds of thousands" of fake accounts to simulate logged-in behavior.

The distinction between survival and collapse is now clear: Courts have generally upheld the right to scrape data that is truly public and non-authenticated (as seen in Bright Data’s successful defense against Meta). However, any company that circumvents security, uses fake accounts, or overrides standard crawling directives—as SerpAPI is accused of doing—is now exposed to existential legal risk.

Proxycurl’s founder, Steven Goh, was candid about the financial reality: without VC funding, the cost of defending against a multi-billion dollar corporation intent on making an example of them was terminal. SerpAPI, which acts as a crucial data pipe for many downstream applications, now faces the same financial gauntlet.

The existential threat to AI search startups

The timing of Google’s lawsuit is critical. It arrives just as a new wave of generative AI search engines—like Perplexity and Exa—are gaining traction by offering summarized, direct answers that often rely on the very structured data that scrapers like SerpAPI extract from Google’s results pages.

Experts suggest that if Google succeeds in crippling SerpAPI, the next targets will be these AI startups themselves. The entire business model of several emerging AI companies is predicated on efficiently accessing and processing the information indexed by Google, often bypassing the traditional click-through model that sends traffic to the original publishers.

Google’s legal action is a powerful, non-technical mechanism to regain control over the data supply chain. By cutting off the source, Google effectively starves its nascent competitors of the raw material they need to function.

This litigation also forecasts a significant shift in how data access is managed online. If scrapers cannot circumvent security measures, the market will inevitably pivot toward regulated, paid access. This suggests a future where systems implemented by companies like Cloudflare and Akamai—which charge on a "pay per bot scrape" model—will become the dominant, legally defensible method for large-scale data collection.

For the thousands of customers who rely on SerpAPI for structured search results, the lawsuit introduces immediate uncertainty and forces a difficult calculus: is the cost savings of using a third-party scraper worth the risk of sudden, legally mandated service termination?

The message from Mountain View is unambiguous: growth and technical prowess alone do not guarantee survival in the data economy. In the new era of AI competition, legal defensibility and the ability to absorb a protracted fight against an incumbent are the only real moats.